The UK’s public broadcaster, BBC, has written a letter to Perplexity, the AI search startup, asking it to stop scraping articles from its websites, delete existing copies of content, and propose some sort of financial compensation if it would like to carry on scraping data. If the demands are not met, BBC may seek an injunction against the startup citing alleged misuse of its intellectual property.
BBC is probably responding in this way because it has seen other news organizations cement deals with firms like OpenAI and Mistral. The income stream allows news organizations to raise more funds and also cover the costs of the extra load on their servers caused by AI scraping.
For anybody not familiar with Perplexity, it’s a bit like ChatGPT but has a much stronger emphasis on searching the web to find information. You can ask it anything you want to know about and it very quickly searches online and constructs a specific response to your question based on what it has found.
The company offers many of its features for free, but does have Perplexity Pro, which costs money. Essentially, Perplexity is making money from publishers by using their content to improve its own product, but not paying them all.
Perplexity's defense and existing publisher programs
In a statement to the Financial Times, Perplexity labeled the BBC’s claims as "manipulative and opportunistic". The startup accused the broadcaster of having “a fundamental misunderstanding of technology, the internet and intellectual property law.”
This is not the first time Perplexity has had a run-in with the media. Forbes and Wired accused it of plagiarizing content from their websites and The New York Times sent the company a cease and desist notice to stop using its content for AI purposes.
To assuage publishers, Perplexity has set up a revenue sharing program, which includes TIME, Fortune, Der Spiegel, and others. According to Digiday, the revenue share was up to 25%. It’s not clear if BBC has tried engaging through this avenue or if it wants to try to squeeze the startup for a bigger slice.
The escalating battle over AI and intellectual property
Even if you only keep up with AI developments in passing, you’ll likely have seen that AI models need to be trained on vast amounts of data, much of which is copyrighted. There is an ongoing debate about whether these companies should be allowed to train on this data, or first seek out permission from the copyright holders.
The move from the BBC could spur other publishers on to try and get themselves a better deal from Perplexity. Alternatively, Perplexity could remove BBC content from its platform and stop pulling information from there. It could probably find most of the information elsewhere, but if Perplexity tried to pull this too much it would eventually end up pretty useless with not a lot of content.
Overall, this is just one of many ongoing legal issues surrounding AI, but once a conclusion has been reached, it could set a precedent about how AI companies should go about getting content from publishers.
6 Comments - Add comment